Meta Expands its Custom Silicon Strategy from Compute into Networking

InfoQ (AI, ML & Data)
Read full post
Meta has introduced MTIA 300, its first custom accelerator designed for training recommendation models, integrating networking directly on-chip to reduce communication bottlenecks. The chip features dedicated message engines and high-bandwidth NICs to enable concurrent compute and communication with minimal performance loss.

More in Chips & Compute

Chips & Compute4 min read

Jensen Huang explains why Nvidia will grow an astounding 70% next year

TechCrunch

Microsoft AI Focused Data Center Plan to Add 26 Gigawatts of Compute

Bloomberg

Tencent-Backed Chipmaker Enflame Jumps 188% in Shanghai Debut

Bloomberg